DESCRIPTION (adapted from the Abstract): This is a revised competing renewal continuation application which is built on research conducted in a multisite, multipurpose study of inpatient AIDS care. The previous study was a comparative, observational study of 40 inpatient units in 20 hospitals in 11 cities in which matching was employed at the hospital level to introduce the control elements of a quasi-experimental design. Data collected included information on organizational attributes of the units and the hospitals, extensive information on nurses, and an array of information on a consecutive sample of AIDS patients including interviews, nurses' clinical assessments, extensive medical record extractions, discharge summaries, posthospital survival status, and billing information. That research effort confirmed a number of positive benefits of dedicated AIDS units over scattered-bed arrangements for AIDS care and offered a number of new and important areas for additional research. In the present study, the Investigator proposes to pursue four new areas: (1) to examine the impact of dedicated AIDS units on care outcomes across hospital type in order to assess the potential of unit level reorganization to improve care in distressed hospitals; (2) to extend the study to a third organizational model--the hospital-wide professional nurse practice model or magnet hospital--which requires additional data collection in six hospitals; (3) to document empirically the operant mechanisms linking variation in hospital organization to patient outcomes; and (4) to estimate the effects of different patient characteristics on the odds on being in a dedicated AIDS unit, and the implications for access to care.